Multicriteria Flood Mitigation in the Imotsko-Bekijsko Polje (Croatia, Bosnia and Herze- Govina)

Home , Polje, Vrljika (river)

DOI: 10.1515/jwld-2015-0018 © Polish Academy of Sciences, Committee for Land Reclamation JOURNAL OF WATER AND LAND DEVELOPMENT and Environmental Engineering in Agriculture, 2015 2015, No. 26 (VII–IX): 73–81 © Institute of Technology and Life Sciences, 2015 PL ISSN 1429–7426 Available (PDF): http://www.itp.edu.pl/wydawnictwo/journal; http://www.degruyter.com/view/j/jwld

Received 15.06.2015 Reviewed 03.08.2015 Accepted 06.08.2015 Multicriteria flood mitigation A – study design B – data collection in the Imotsko-Bekijsko Polje C – statistical analysis D – data interpretation E – manuscript preparation (Croatia, Bosnia and Herzegovina) F – literature search

Igor LJUBENKOV ABCDEF

Water Development Ltd., Split, Croatia, e-mail: [email protected]

For citation: Ljubenkov I. 2015. Multicriteria flood mitigation in the Imotsko-Bekijsko Polje (Croatia, Bosnia and Herze- govina). Journal of Water and Land Development. No. 26 p. 73–81. Abstract Imotsko-Bekijsko Polje has an area of 9 500 ha and is one of the biggest karst fields (polje) in the Dinaric Mountains, extending over the territory of two states: Croatia and Bosnia and Herzegovina. Many hydraulic structures (reservoirs, retentions, tunnels, etc.) have been built since the middle of 20th century in order to protect polje against floods. Therefore, the security from flooding has increased substantially. However, there is still periodical flooding in the southeastern lowest part of the polje. The largest flood in recent times was in January 2010, when 2676 ha (28% of the area) was flooded. The polje is a typical karst with very complex hydrological and hydrogeological relations. In this paper two hydrological stations, Nuga at the lowest part and Kamenmost in the central part of the polje with respectable hydrological series, are statistically analysed. In particular, the effi- ciency of existing hydraulic structures for flood mitigation is estimated. The research points out that floods in Imotsko-Bekijsko Polje are largely influenced by water management objects (reservoir, retention, tunnel) and only indirectly by precipitation.

Key words: flood, Imotsko-Bekijsko Polje, karst polje, statistical test

INTRODUCTION

Dinaric karst is an extent, 800 km long and up to 150 km wide area of karst, which extends over majority of the Dinaric Mountains on the Balkan peninsula in SE Europe and through the western and southern part of Croatia (regions Istria, Kvarner, Lika and Dalmatia) and the majority of Bosnia and Herzego- vina (Fig. 1) [BONACCI 1987; MIHEVC et al. 2010]. The polje in the karst area can be defined as an extensive depression in carbonate massif, with fertile soil and relatively mild slope. Karst fields usually have numerous hydrological and hydrogeological forms such as permanent and temporary springs, streams, ponors, estavelles, etc. with a very complex and variable hydrological relationships [BONACCI 1987; 2004; BONACCI, LJUBENKOV 2006; KRESIC 2013; MIJATOVIĆ 1988; SACKL et al. 2014]. Fig. 1. Map of the Dinaric karst; source: own elaboration

In the Dinaric karst, fields are restricted areas river regulation. However, intensive flood control and with the most favourable conditions for agriculture construction of appropriate facilities (reservoirs, re- and life. The majority of those fields are elongated in tention reservoirs, tunnels, dams, canals and pumping one direction (NW–SE), surrounded by bare and stations) began in mid-20th century, after which the sometimes inaccessible rocky terrain. Although rela- general state of the Imotsko-Bekijsko Polje has been tively small in area, these Dinaric fields have enor- improved. The aim of this paper is to describe the mous social and economic role. They have always regime of high water in Imotsko-Bekijsko Polje and been used for agriculture, and partly for animal hus- to give an overview of the role of the existing facili- bandry. ties in flood mitigation. In Dalmatia, southern Croatian region, there are more than one hundred karst poljes with area larger HYDROLOGICAL REGIME than 10 ha. Eleven of them has an area of more than 1 000 ha and the largest is Imotsko-Bekijsko Polje STUDY AREA (9 500 ha). Almost all of them are subject to occasional flooding in the cold and wet season (October to Imotsko-Bekijsko, the largest Dalmatian karst 2 April), although some of them are meliorated. polje has an area of 9500 hectares (i.e. 95 km ), of SCHWARZ [2013] summarised the potential flooding which 4400 hectares is located in the Republic of situation for 57 poljes in Bosnia and Herzegovina Croatia, and the remaining 5100 hectares in neigh- (B&H), including Imotsko-Bekijsko Polje. His ap- bouring Bosnia and Herzegovina (Fig. 2). The higher proach has been based on elevation model but he also western part of the polje (altitude between 254 and took into account flood documentation, historical 280 m a.s.l.) belongs to Croatia while the eastern low- maps, etc. er part (altitude between 249 and 266 m a.s.l.) belongs Since floods make heavy damage, especially in to B&H. The topographic catchment of the polje cov- 2 agriculture, man has always exercised a number of ers 325 km . Besides from direct catchment, Imotsko- interventions in poljes and their catchment in order to Bekijsko Polje receives surface water from the Suvaja improve the hydrological regime i.e. to reduce the (Ricina) River on the north-western edge of the polje spatial extent and duration of floods. This problem is (Fig. 2). The Ricina collects water in the area of Po- interesting also from the scientific point of view. For susje and brings it to the reservoir Ricice. The topog- 2 example, BONACCI [1987] analysed the impact of hy- raphic catchment of the Ricina is about 300 km , lo- draulic structures on flood control in Vrgorsko Polje cated at higher areas between 300 and 1300 m a.s.l. (Croatia). ZHANG et al. [2014] applied t and F test and Obviously, it is an indirect catchment of the polje. determined abrupt streamflow changes of the East Downstream the reservoir Ricice this watercourse is River in China influenced by both water reservoir called the Suvaja. It flows along a 3.8 km canyon construction and precipitation changes. through the polje and finally it finishes in Prolosko The first land-improvement works in the area of blato. The polje is also fed by waters from numerous Imotsko-Bekijsko Polje took place in the 19th century karst springs located on the northern edge of the polje and involved increasing the capacity of ponors and (Opacac, Krenica, Slavic, Grudsko vrilo etc.). So, water comes to polje from the higher area of Posusje i.e. an indirect part of the catchment by surface flow (the Suvaja) as well as from karst underground (springs). Climate of this area is influenced by the Mediterranean Sea from the west and by the continen- tal hinterland. Average annual precipitation at the meteorological station Imotski (43°27' N, 17°13' E, 435 m a.s.l.) was 1 280 mm in the period 1957–2013. Imotski is the biggest place in this area. The minimum annual precipitation was 746 mm (year 1983) while the maximum was 1905 mm (year 2010). The summer months could be without precipitations. The highest precipitations are in the autumn-winter period. The rainiest month is November with an aver- Fig. 2. Map of catchment area of the Imotsko-Bekijsko Polje; age precipitation of 189 mm. The source: own elaboration average annual temperature is

13.8°C. The coldest month is January with 5.0°C and reservoirs were built, Ricice in 1989, Tribistovo the hottest is July with 24.0°C. (1990) and IGM (1994), situated in the “indirect” part The entire catchment area of the polje has very of the polje's catchment. The construction of major complex hydrogeological relations [BONACCI et al. facilities in this water management system ended in 2013]. Certainly, topographic (morphological) and 2004 when HEPP Pec Mlini started to work. Its sup- hydrogeological catchment boundary of Imotsko- ply tunnel (capacity 30 m3·s–1) was set parallel to the -Bekijsko Polje are different. The real boundaries of tunnel Petnjik. Therefore, drainage of the lowest area the catchment are mostly underground and rarely to- of the polje is nowadays achieved via tunnel Petnjik pographic. and HEPP's tunnel. The polje is flooded generally in The catchment area of the polje is built of Creta- wet periods but flooding depends also on human ac- ceous limestone and dolomite, sandstone, shale and tivity i.e. on the operation of hydraulic structures. For conglomerate, Miocene marl and clay limestone and example, in the flood risk period along the Tihaljina, Quaternary alluvial and talus deposits. Carbonate de- tunnels are partially or fully closed. Therefore, it in- posits are very tectonically fractured and highly kar- creases water level in the retention reservoir Nuga stified and represent all types of karst phenomena. contributing flood to the lowest part of the polje. Quaternary sediments (Q) prevail in the field with Reservoir Ricice was planned as a multi-purpose alluvial (al) sandstones, clay sands and gravels and object for flood control, irrigation and water supply. It their mixtures (limestone debris with clay and red is the biggest reservoir in this area. Unfortunately, it soil). The thickness of these layers varies and reaches did not meet expectations in terms of providing suffi- more than 130 m. cient quantities of water for irrigation of the polje. It The only permanent watercourse, the Vrljika was designed to a volume of 35.2 million m3, but it River of a length of about 18 km, flows from its has never been completely filled. This is due to leach- source Opacac to the lowest area Nuga, meandering ing losses from the reservoir itself, but also to reduced through the polje. Besides, the polje is cut with nu- inflows from the catchment Ricina compared with the merous drainage and irrigation canals. project documentation. However, after the construc- Since, the Imotsko-Bekijsko Polje and associated tion of reservoir Ricice (1989), high surface water water management facilities are placed in the territory waves that reach Imotsko-Bekijsko Polje have been of two countries, it is so-called “intersected” system. reduced and thus polje's flood protection has been It is actually a natural and technical unit but separated improved. in two parts from administrative point of view (Fig. Downstream the reservoir Ricice, surface water 2). Also, Imotsko-Bekijsko Polje is higher and up- flows through the Suvaja riverbed, whose discharge stream area for a few lower downstream poljes: largely depends on the releases from the reservoir, Ljubuško Polje (B&H), Rastok (B&H and Croatia) except in the extreme flood situations. Now, this wa- and Vrgorsko Polje (Croatia), located between 30 and tercourse is mainly used for releasing water for irriga- 75 m a.s.l., so it has an important hydrological influ- tion. Through the Suvaja riverbed, surface waters en- ence on them. ter the polje in Prolozac where they are partially di- Under natural conditions, Imotsko-Bekijsko is verted in irrigation canal while most of water contin- drained only through ponors located in southern edge ues its flow to retention reservoir Prolosko blato con- of the polje (ponors Nuga, Kongora, Prispa etc.). structed in 1956. Water from this reservoir is dis- There was a flood when capacity of ponors was insuf- charged into Sija canal through sluice gate. Canal is ficient. The area most vulnerable to flooding was the 7.2 kilometres long and flows into the Vrljika River. southern part of the field (Nuga), which is the lowest After the construction of the Ricica dam, Sija canal is one (249 m a.s.l.), so all surface flow runs towards it. mostly empty. In the previous period, about 120 km of drainage FLOOD CONTROL OBJECTS canals were constructed in the polje apart from men- tioned above facilities. Also, the most important wa- Tunnel Petnjik is the first built, and also the most tercourse the Vrljika was regulated. Therefore, drain- important object for flood mitigation in the Imotsko- age network of polje is solid though it requires inten- -Bekijsko Polje (Fig. 2). Construction of the tunnel sive maintenance (cutting of vegetation, removal of began before World War II, and was completed and debris etc.). put into operation in 1951. The capacity of the tunnel is about 40 m3·s–1. The tunnel serves for the evacua- HIGH WATER REGIME tion of water from Nuga to the lower area i.e. watercourse Tihaljina, which continues its course under the Hydrological conditions in the Imotsko-Bekijsko name Mlada and Trebizat until its inflow into the Polje are monitored in six locations: Prolozac (at the Neretva River. Further on, the Neretva River flows Suvaja and irrigation canal), Opacac (at the Vrljika into the Adriatic Sea. and irrigation canal), Kamenmost (at the Vrljika) and Then, two retention reservoirs Rastovaca on the the Sija (the homonymous waterourse) in Croatia and Ricina and Prološko blato on the western edge of the Grudsko vrilo and Nuga in B&H. Hydrological sta- polje were completed in 1956. Furthermore, three tions at Suvaja and Opacac have relatively short

© PAN in Warsaw, 2015; © ITP in Falenty, 2015; Journal of Water and Land Development. No. 26 (VII–IX) 76 I. LJUBENKOV measurement series (less than 20 years), so they are not acceptable for statistical analysis. According to engineering practice, hydrological time series should be at least 30 years long to apply statistical methods [PROHASKA 2003; RAGHUNATH 2006]. Furthermore, gauging station Sija has sufficiently long measure- ments but it is directly influenced by retention reser-

1 voir Prolosko blato, therefore, it was not considered in – ·s this paper. Station Grudsko vrilo has a relatively long 3 series, but it could not register the highest flow rates. , m Q So it is not useful for flood analysis. The remaining two stations, Kamenmost and Nuga have a relatively long series (N = 56 and 72, respectively) so the appropriate statistical analysis could be applied to them. They are also the most important stations in this area from the hydrological point of view. Kamenmost closes the whole indirect part of the catchment as well as western part of the polje and covers the majority of water that causes flooding. Nuga is in the lowest part of the area and is critical for floods. The methods used in this work: t-test and F-test are widely used, com- monly accepted and can be applied to any other simi- lar sequences. These statistical tests are given in many

papers [BONACCI 1987; MIRZA 2004; ZHANG et al. – ·s 2014]. BONACCI [1987] used the same approach for 3

Vrgorsko Polje analysis. , m The Vrljika River flows are detected at the hydro- Q logic station Kamenmost, which has been in operation since the end of the 19th century. But the station in this long period was shifted and reconstructed several times. Limnigraph was installed in 1956, since then it registered discharges. It measures flow from the entire upstream catchment except a relatively small amounts Fig. 3. Characteristic monthly discharges in Kamenmost: of water that flows through the main irrigation canal a) in the operation period, b) after Ricice reservoir; Qmax = located approximately parallel to the Vrljika, which is maximum recorded discharge, Qmaxav = mean maximum used only in the vegetation period (from April to Sep- monthly discharge, Qav = mean monthly tember). Fig. 3a shows maximum discharges, mean discharge; source: own study high and monthly mean discharges in Kamenmost in the operation period 1957–2013. The highest flow of polje, next to the entrance to the drainage tunnel and 99.4 m3·s–1 was recorded in December 1959. Mean has a key role in flood monitoring. Water levels have monthly flows were less than 20 m3·s–1 and could not been recorded since 1926 with interruptions in 1943, cause floods. Discharges greater than 50–60 m3·s–1 1944, from 1946 to 1949, 1992, 1993, 1994 and from can cause flood locally along the Vrljika as well as in 2003 to 2005. Detailed analysis of high water in this the lowest area of Nuga. After the construction of location is presented below. Ricice retention reservoir, maximum discharges were Another gauging station in Bosnia and Herzego- reduced. Fig. 3b shows characteristic monthly flows vina is Grudsko vrilo, located in springs of the same in new conditions (period 1989–2013) that reflect the name. Water from this source flows to the retention actual status. The maximum flow recorded in Decem- reservoir Nuga by a drainage canal. The station began ber 2004 was 65.5 m3·s–1. Obviously, flood risk was operation in 1961, with an interruption of work on reduced but floods are still possible. We can say that several occasions. The highest flow rates in this pro- water flow in this profile after 1989 has two compo- file can not be determined. nents – the artificial and the natural. Artificial one arises from the working regime of hydraulic structures RESULTS such as the reservoir Ricice (1989) and Prolosko blato (1956), while the natural flow depends on the source Before construction of the tunnel Petnjik (1951), Opacac and surface inflow in wet period. Detailed the field was flooded almost every year. The largest analysis of high water recorded in this gauging loca- recorded flood before construction of the tunnel was tion is explained below. in winter 1934/1935 when the flood covered 5333 ha Nuga gauging station is located in the retention (56% of the field), and lasted approximately 6.5 reservoir of the same name in the lowest part of the months (Figs. 4, 5 and 6). In the following sub-period

hmax

Fig. 4. Series of maximum annual water levels in Nuga; A: before tunnel Petnjik construction, B: after tunnel construction, C: after reservoir Ricice construction, D: present conditions i.e. after HEPP Pec Mlini construction; source: own study

Fig. 5. Series of flooded area in the period 1927–2013; A–D as in Fig. 4; source: own study

Fig. 6. Comparison of flood events in different periods; A: before tunnel Petnjik construction, B: after tunnel construction, C: present conditions i.e. after HEPP Pec Mlini construction; source: own study

(from 1951 to 1989), after construction of the tunnel Table 1. Data on high water levels in Nuga in various periods Petnjik but before construction of the Ricice retention 1927– 1951– 1988– 2004– Parameter reservoir, the largest flood was recorded in winter 1951 1988 2004 2014 1959/1960. Its maximum coverage was 4097 ha (43% No. of data 16 38 9 9 of the field), and total duration was 89 days. In the Average high water 256.83 253.88 253.62 254.71 next period (1989/1990–2003/2004), after construc- level, m a.s.l. Maximum recorded tion of the reservoir Ricica and before HEPP Pec 260.18 258.58 255.91 256.74 Mlini, the maximum water elevation was 255.91 m water level, m a.s.l. a.s.l. registered in January 1996. Corresponding hmax date Mar 1935 Dec 1959 Jan 1996 Jan 2010 flooded area was 2035 ha. Later, after HEPP con- Variance 4.48 5.41 5.91 3.16 struction, the largest flood was in January 2010 with Skewness –0.07 –0.03 –0.01 –0.77 water table elevation of 256.74 m a.s.l. It flooded Explanations: hmax = maximum recorded water level. Source: own study. 2676 hectares (28% of the field), and the corresponding amount of water was 57·106 m3. The duration of means are equal was accepted. For example, construc- the flood was only 23 days. tion of Prolosko blato retention reservoir as well as Fig. 5 shows the size of flooded areas in ha, that HEPP Pec Mlini one did not statistically change high corresponds to maximum water elevation per hydro- water mean (Tabs. 3, 4). The increase of high water logical years from 1927/1928 to 2012/2013. Unfortu- mean is evident only for HEPP (0.09 m) while tunnel nately, there were interruptions in water observations and retention reservoir decreased it. There were no but qualitative explanation of flood processes is significant differences of variances among these sub- achieved due to statistical analysis. Essential data for periods (Tab. 2, 3 4). Therefore, the variability of high analysed periods are given in Tab. 1. Maximum water water regime did not change despite the construction tables from different sub-periods are compared in Fig. of hydrotechnical objects. 6. It is evident that floods are now mitigated i.e. flood The two sub-periods were analysed in Kamen- extent and duration are reduced. most, before and after construction of Ricice retention There was a significant (α = 0.05) difference be- resevoir (Fig. 7). Basic hydrologic data are presented tween mean water levels at station Nuga in two time in Table 5. Construction of this reservoir did not sig- series: 1927/1928–1951/1952 and 1952/1953 to nificantly change water peaks (Q ) and its variabil- 1988/1989, i.e. before and after the construction of the max ity in hydrological series (Tab. 6). The largest water tunnel Petnjik (Tab. 2). In this case statistical hy- flow from the second sub-period (1989/1990– pothesis H was accepted. Other periods did not show 1 2012/2013) had a peak flow rate of 65.5 m3 s–1 (Dec significant differences, so H hypothesis that the two 0 2004).

Table 2. Statistical tests for Nuga: the influence of tunnel Petnjik and Ricica reservoir

Location: Nuga Parameter: Water table elevation 2 hav 2 h α = 0.05 σ α = 0.05 No. period Ni s hypothesis hypothesis m a.s.l. t-test tcrit F-test Fcrit 1 1927/1928–1950/1951 16 256.83 4.48 2.27 2.01 H1 0.83 0.45 H0 2 1951/1952–1988/1989 38 253.88 5.41

0.34 2.01 H0 0.87 0.52 H0 3 1989/1990–2012/2013 18 253.60 6.19 2 Explanations: Ni = No. of data, hav = mean high water level, s = high water level variance. Source: own study.

Table 3. Statistical tests for Nuga: the influence of HEPP Pec-Mlini

Location: Nuga Parameter: Water table elevation 2 hav 2 h α = 0.05 σ α = 0.05 No. period Ni s hypothesis hypothesis m a.s.l. t-test tcrit F-test Fcrit 1 1951/1952-2001/2002 47 253.62 5.91 –1.17 2.01 H0 1.87 3.01 H0 2 2005/2006-2012/2013 9 254.71 3.16 Explanations as in Tab. 2. Source: own study.

Table 4. Statistical tests for Nuga: the influence of Prolosko blato

Location: Nuga Parameter: Water table elevation 2 hav 2 h α = 0.05 σ α = 0.05 No. period Ni s hypothesis hypothesis m a.s.l. t-test tcrit F-test Fcrit 1 1951/1952–1955/1956 5 255.40 1.77 0.38 2.01 H0 0.31 0.18 H0 2 1956/1957–2012/2013 51 253.63 5.70 Explanations as in Tab. 2. Source: own study.

Fig. 7. Series of maximum annual discharges in Kamenmost; A: before reservoir Ricice construction, B: after reservoir construction; source: own study

Table 5. Data for maximum discharges in Kamenmost in Special attention in this work was paid to analyse different periods the effect of precipitation on floods. Therefore, pre- Parameter 1957–1988 1989–2013 cipitation data were taken from the nearest meteoro- No. of data 32 24 logical station Imotski for the period 1957–2013. The Mean annual high water discharge, Mann–Kendall test [ZHANG et al. 2014] showed sig- 54.5 43.0 m3·s–1 nificantly decreasing precipitation on annual scale Maximum recorded water dis- 99.4 65.5 (about –3 mm per year) and decreasing monthly max- charge, m3·s–1 ima (about –1.4 mm per year, Fig. 8). Fig. 8 shows Qmax date Dec 1959 Dec 2004 the relationship between monthly maximum precipita- Variance 320.9 220.9 tion and corresponding maximum discharge (the Skewness 0.65 - 0.33 Vrljika, station Kamenmost) registered in the same Explanation: Qmax = maximum recorded water discharge. month. There was a weak correlation between precipi- Source: own study. Table 6. Statistical test for Kamenmost: the influence of Ricice reservoir

Location: Nuga Parameter: Discharge 2 Qav 2 Q α = 0.05 σ α = 0.05 No. period Ni 3 –1 s hypothesis hypothesis m ·s t-test tcrit F-test Fcrit 1 1957/1958–1988/1989 32 54.5 320 1.90 2.01 H0 1.45 1.96 H0 2 1989/1990–2012/2013 24 43.0 220 2 Explanations: Ni = No. of data, Qav = mean high water discharge, s = high water discharge variance. Source: own study.

Linear trend (P)

Fig. 8. Series of maximum monthly precipitations Pmax with corresponding maximum discharges Q; source: own study

1 1 – – , mm·month , mm·month max max y = 0.0444x + 244.93 P P y = 1.976x + 188.94 R2 = 0.3344 R2 = 0.3107

Q, m3·s–1 F, ha

Fig. 9. Correlation between maximum monthly precipitations Pmax and corresponding discharges Q (a) and flooded area F (b); source: own study tation and flow in the Vrljika (Q) and between pre- by reducing water peaks and decreasing risk of water cipitation and respective flooded area (F) (Fig. 9). For spreading from the Vrljika riverbed. example, the maximum monthly precipitation (443 Despite previous engineering engagement, agri- mm) registered in December 1959 generated the max- cultural land in the south-eastern lowest part of the imum surface flow (99.4 m3·s–1) in the polje and max- polje is still subject to occasional flooding in the win- imum flood (4 097 ha). But the same monthly precipi- ter months. The Petnjik tunnel has sufficient capacity tation in October 1974 (443 mm) generated surface for high water evacuation. But occasionally it has to flow of 60.9 m3·s–1 without flood. These relationships be closed completely or partially because of down- illustrate the natural complexity of karst environment stream runoff conditions, which is the crucial problem [BONACCI et al. 2013], even if based on monthly scale for flood control. The solution of this problem re- as in this work. quires engagement of institutions and experts from the In comparison to the research of ZHANG et al. two countries, Croatia and B&H, because it is a bor- [2014], streamflows and floods in Imotsko-Bekijsko der area. Polje are largely influenced by water management objects (reservoirs, retention, tunnel) and only indi- REFERENCES rectly by precipitation. BONACCI O. 1987. Karst hydrology: With special references CONCLUSION to the Dinaric karst. Berlin. Springer Verl. ISBN 3540181059 pp. 184. Since the middle of the 20th century, numerous BONACCI O. 2004. Poljes. In: Encyclopedia of caves and karst science. Ed. J. Gunn. New York. Fitzroy Degrborn facilities have been constructed for water use and p. 559–600. flood control in the Imotsko-Bekijsko Polje, including BONACCI O., LJUBENKOV I. 2006. Karst flash floods: an Ricina catchment, from where water comes to the example from the Dinaric karst (Croatia). Natural Haz- polje by surface and underground flow. Past activities ards and Earth System Science. Vol. 6 p. 195–203. significantly improved the situation, through increas- BONACCI O., ŽELJKOVIĆ I., GALIĆ A. 2013. Karst river’s ing the safety from flood, decreasing damage to agri- particularities: an example from the Dinaric karst (Croa- culture and water management, and providing a con- tia/Bosnia and Herzegovina). Environmental Earth Sci- siderable reserve of water for water supply and irriga- ence. Vol. 70 p. 963–974. tion. KRESIC N. 2013. Water in karst: Management, vulnerability and restoration. New York. McGraw-Hill. ISBN Obviously, the most important object for flood 978071753333 pp. 708. control is the tunnel Petnjik and then Ricica reservoir. MIHEVC A., PRELOVŠEK M., HAJNA N.Z. 2010. Introduction However, the tunnel Petnjik has an influence on the to the Dinaric karst. Postojna. Karst Research Institute. lowest part of the polje, which is most vulnerable in ISBN 9789612541989 pp. 71. the present conditions. The comparison of different MIJATOVIĆ B. 1988. Catastrophic flood in the polje of Cet- water waves (Fig. 6) shows that flood control in the inje in February 1986, a typical example of the envi- Imotsko-Bekijsko Polje has been increased due to ronmental impact of karst. Environmental Geology and human activity. The Ricice reservoir and Prolosko Water Sciences. Vol. 12. Iss. 2 p. 117–121. blato retention reservoir do not affect the floods of MIRZA M.M.Q. 2004. The Ganges water diversion: environmental effects and implications. Dordrecht. Kluwer lower area (Nuga), but they have an impact on the Acad. Publ. ISBN 9789048166657 pp. 367. flood control in the middle and upper part of the polje

PROHASKA S.J. 2003. Hidrologija [Hydrology]. Beograd. SCHWARZ U. 2013. Flooding analysis of karst poljes in Bos- Rudarsko-geološkifakultet pp. 428. nia & Herzegovina. Vienna. FLUVIUS pp. 127. RAGHUNATH H.M. 2006. Hydrology: principles, analysis, ZHANG Q., SINGH V.P., LI K., LI J. 2014. Trend, periodicity design. New Delhi. New Age Int. ISBN 8122418252 pp. and abrupt change in streamflow of the East River, the 476. Pearl River basin. Hydrological Processes. Vol. 28. Iss. SACKL P., DURST R., KOTROŠAN D., STUMBERGER B. 2014. 2 p. 305–314. Dinaric Karst Poljes – Floods for life. Radolfzell. EuroNatur. ISBN 978-3-00045287-1 pp. 199.

Igor LJUBENKOV

Ochrona przed powodzią w Dolinie Imotsko-Bekijsko Polje (Chorwacja, Bośnia i Hercegowina)

STRESZCZENIE

Słowa kluczowe: Imotsko-Bekijsko Polje, powódź, tereny krasowe, testy statystyczne Imotsko-Bekijsko Polje ma powierzchnię 9500 ha i jest jedną z największych dolin krasowych w Górach Dynarskich. Teren doliny rozciąga się na terytorium dwóch państw – Chorwacji oraz Bośni i Hercegowiny. Budowle hydrotechniczne (zbiorniki retencyjne, tunele itp.) zostały zbudowane w połowie XX wieku w ce- lu ochrony doliny przed powodziami. W związku z tym bezpieczeństwo powodziowe znacznie wzrosło, jednak występują jeszcze okresowe powodzie w południowo-wschodniej, najniższej części doliny. Największa powódź w ostatnim czasie zdarzyła się w styczniu 2010 r., zalane zostało wówczas ok. 2676 ha, co stanowi 28% powierzchni terenu. Polje jest typowo krasowym terenem o bardzo złożonymi stosunkami hydrologicznymi i hydrogeologicz- nymi. W niniejszej pracy analizie statystycznej poddano dane z dwóch stacji hydrologicznych – Nuga w najniż- szej części i Kamenmost w środkowej części doliny. Szczególnie analizowano efektywność istniejących struktur hydrologicznych dla łagodzenia skutków powodzi. Wyniki badań wskazują, że powodzie w Imotsko-Bekijsko Polje wywoływane są głównie przez obiekty gospodarki wodnej (zbiorniki retencyjne, tunele itp.) i tylko po- średnio zależą od opadów atmosferycznych.